This activity aims at developing the technical backbone of BloomingAI. The team will fine‑tune the AI model for educational use, define its system requirements, and deliver a secure, web‑based application that follows the UX patterns defined in the work package “Mapping State-of-the-Art Practices in AI-Assisted Learning”. This includes optimizing servers, software, data processing pipelines, and user interfaces to ensure a smooth, scalable, and GDPR-compliant deployment of BloomingAI in higher‑education settings.

We will develop a strategy for making BloomingAI open source, allowing educational institutions and developers to adapt and extend its functionalities and ensuring long-term sustainability and accessibility. This includes creating a strategy for open-access implementation, defining licensing, documentation, contribution pathways, and sustainability measures so institutions and researchers can adopt and extend the platform. By making the platform open-source, the project ensures adaptability, modifiability, and continuous refinement based on user needs.

BloomingAI will be tested across all partner universities, in different learning environments, to evaluate usability, pedagogical effectiveness, learning impact, curricular fit, and alignment with learning objectives. Educators and students will use the assistants in real courses, and using the standardized feedback toolkit developed by our team we will collect feedback, and analyse the platform’s effectiveness in teaching, learning, and assessment. Our findings will drive iterative improvements to both the platform and the integration approach.  

The University of Luxembourg will organize a workshop bringing together partners, educators, and AI researchers to consolidate insights from the piloting phase. The workshop will facilitate discussions on best practices, challenges, recommendations, translating user evidence into concrete refinements. The outcome will be a validated set of improvements and clear guidance for broader institutional adoption.